Reproduce “CrossTL: Hub-Based Translation of GPU Programs and the Preservation of Source Rules” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

NRIPESH NIKETAN, VAATSALYA SHRIVASTVA, BHUSHAN SRIVASTAVA, SUSHANT MISHRA

CrossTL-R records selected source-operation rules in a translation hub and lowers them through target-specific entries, using native operations, polyfills, or refusal. The paper tests this approach on a 24-case hazard comparison: CrossTL has 37 divergences, local SPIR-V fixes have 23, and CrossTL-R has 5. All five remaining failures occur on MoltenVK, so the results do not establish correctness across runtime stacks.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.